Speed up some test by reducing deconvolution iterations

[anawas]

This PR closes the issue #647. Lowering the number of iterations speeds up some of the tests. Running them on my machine with RUN_NOTEBOOK_TESTS=true pytest --durations=10 gives:

205.05s call     karabo/test/test_notebooks.py::test_imaging_notebook
184.72s call     karabo/test/test_imager_rascil.py::test_create_cleaned_image
174.71s call     karabo/test/test_image.py::test_imaging
159.13s call     karabo/test/test_notebooks.py::test_source_detection_notebook
80.54s call     karabo/test/test_line_emission.py::test_primary_beam_effects[SimulatorBackend.RASCIL-MID]
56.45s call     karabo/test/test_line_emission.py::test_primary_beam_effects[SimulatorBackend.OSKAR-SKA1MID]
51.59s call     karabo/test/test_skymodel.py::test_read_healpix_map
35.81s call     karabo/test/test_mock_mightee.py::test_mock_mightee
26.18s call     karabo/test/test_source_detection.py::test_full_source_detection
23.77s call     karabo/test/test_notebooks.py::test_source_detection_big_files_notebook

I also rewrote some other test cases in order to make them more meaningful.